1 / 50

Instructor: Justin Hsia

CS 61C: Great Ideas in Computer Architecture Memory Management and Usage. Instructor: Justin Hsia. Review of Last Lecture. Arrays Traverse using pointer or array syntax If characters, call them “strings” and null-terminate

yosefu
Download Presentation

Instructor: Justin Hsia

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CS 61C: Great Ideas in Computer Architecture Memory Management and Usage Instructor: Justin Hsia Summer 2012 -- Lecture #4

  2. Review of Last Lecture • Arrays • Traverse using pointer or array syntax • If characters, call them “strings” and null-terminate • Pointer arithmetic moves the pointer by the size of the thing it’s pointing to • C accommodates for pointing to structs and pointers • From Discussion: Bit-masking Summer 2012 -- Lecture #4

  3. lw $t0, 0($2) lw $t1, 4($2) sw $t1, 0($2) sw $t0, 4($2) Great Idea #1: Levels of Representation/Interpretation We are here_ temp = v[k]; v[k] = v[k+1]; v[k+1] = temp; Higher-Level LanguageProgram (e.g. C) Compiler Assembly Language Program (e.g. MIPS) Assembler 0000 1001 1100 0110 1010 1111 0101 1000 1010 1111 0101 1000 0000 1001 1100 0110 1100 0110 1010 1111 0101 1000 0000 1001 0101 1000 0000 1001 1100 0110 1010 1111 Machine Language Program (MIPS) Machine Interpretation Hardware Architecture Description(e.g. block diagrams) Architecture Implementation Logic Circuit Description(Circuit Schematic Diagrams) Summer 2012 -- Lecture #4

  4. Agenda • C Memory Layout • Stack, Static Data, and Code • Administrivia • Dynamic Memory Allocation • Heap • Memory Management • C Wrap-up: Linked List Example • Bonus: Common Memory Problems Summer 2012 -- Lecture #4

  5. C Memory Layout ~ FFFF FFFFhex stack • Program’s address spacecontains 4 regions: • Stack: local variables, grows downward • Heap: space requested for pointers via malloc(); resizes dynamically, grows upward • Static Data: global and static variables, does not grow or shrink • Code: loaded when program starts, does not change heap static data code ~ 0hex OS prevents accessesbetween stack and heap (via virtual memory) Summer 2012 -- Lecture #4

  6. Where Do the Variables Go? • Declared outside a procedure: Static Data • Declared inside a procedure: Stack • main() is a procedure • Freed when procedure returns • Dynamically allocated: Heap • i.e. malloc (we will cover this shortly) #include <stdio.h> intvarGlobal; int main() { intvarProc; int *varDyn = malloc(sizeof(int)); } Summer 2012 -- Lecture #4

  7. frame frame frame The Stack Function call: Function returns: • Each stack frame is a contiguous block of memory holding the local variables of a single procedure • A stack frame includes: • Location of caller function • Procedure arguments • Space for local variables • Stack pointer (SP) tells where lowest (current) stack frame is • When procedure ends, stack frame is tossed off the stack; frees memory for future stack frames frame SP … Summer 2012 -- Lecture #4

  8. The Stack stack • Last In, First Out (LIFO) data structure int main() { a(0); return 1; } void a(int m) { b(1); } void b(int n) { c(2); e(4); } void c(int o) { printf(“c”); } void d(int p) { printf(“d”); } Stack grows down Stack Pointer Summer 2012 -- Lecture #4

  9. SP SP SP getPtr()(y==3) printf()(y==?) main main main Stack Misuse Example Neverreturn pointers to local variable from functions Your compiler will warn you about this – don’t ignore such warnings! int *getPtr() {inty; y = 3; return &y;}; int main () {int*stackAddr,content; stackAddr= getPtr(); content = *stackAddr;printf("%d", content); /* 3 */ content = *stackAddr;printf("%d", content); /* 0 */ }; What’s BAD about this function? stackAddr printf overwrites stack frame Summer 2012 -- Lecture #4

  10. Static Data • Place for variables that persist • Data not subject to comings and goings like function calls • Examples: String literals, global variables • Size does not change, but its data can Code • Copy of your code goes here • C code becomes data too! • Does not change Summer 2012 -- Lecture #4

  11. Question: Which statement below is FALSE?All statements assume each variable exists. void funcA() {int x; printf(“A”);} void funcB() { int y; printf(“B”); funcA(); } void main() {char *s = “s”; funcB();} x is at a lower address than y ☐ x and y are in adjacent frames ☐ x is at a lower address than *s ☐ y is in the 2nd frame from the top of the Stack ☐

  12. Agenda • C Memory Layout • Stack, Static Data, and Code • Administrivia • Dynamic Memory Allocation • Heap • Memory Management • C Wrap-up: Linked List Example • Bonus: Common Memory Problems Summer 2012 -- Lecture #4

  13. Administrivia • Meet your fellow classmates! Form study groups and get your questions answered • Utilize Piazza, labs, discussions, and OHs • End of the first week! • HW1 due Sunday • TA and Instructor response time increase drastically (slower) over the weekend • Check key card access to lab today • Justin’s OH (this week only): Sat @ 10-12 in lab • Suggestion for weekend: finish HW1, try Lab3, then look at HW2Q1 Summer 2012 -- Lecture #4

  14. Agenda • C Memory Layout • Stack, Static Data, and Code • Administrivia • Dynamic Memory Allocation • Heap • Memory Management • C Wrap-up: Linked List Example • Bonus: Common Memory Problems Summer 2012 -- Lecture #4

  15. Dynamic Memory Allocation • Sometimes you don’t know how much memory you need beforehand • Examples: Input files, user input • Dynamically allocated memory goes on the Heap – more permanent than Stack • Need as much space as possible without interfering with Stack • Start at opposite end and grow towards Stack Summer 2012 -- Lecture #4

  16. Allocating Memory • 3 functions for requesting memory:malloc(), calloc(), and realloc() • http://en.wikipedia.org/wiki/C_dynamic_memory_allocation#Overview_of_functions • malloc(n) • Allocates a continuous block of n bytes of uninitialized memory (contains garbage!) • Returns a pointer to the beginning of the allocated block; NULL indicates failed request (check for this!) • Different blocks not necessarily adjacent Summer 2012 -- Lecture #4

  17. Using malloc() • Almost always used for arrays or structs • Good practice to use sizeof() and typecasting int *p = (int *) malloc(n*sizeof(int)); • sizeof() makes code more portable • malloc() returns void *, typecast will help you catch coding errors when pointer types don’t match • Can use array or pointer syntax to access • Make sure you don’t lose the original address • p++ is a BAD IDEA; use a separate pointer Summer 2012 -- Lecture #4

  18. Releasing Memory • Release memory on the Heap using free() • Memory is limited, release when done • free(p) • Pass it pointer p to beginning of allocated block; releases the whole block • p must be the address originally returned by m/c/realloc(), otherwise throws system exception • Don’t call free() on a block that has already been released Summer 2012 -- Lecture #4

  19. Dynamic Memory Example • Need #include <stdlib.h> typedefstruct { int x; int y; } point; point *rect; /* opposite corners = rectangle */ ... if( !(rect=(point *) malloc(2*sizeof(point))) ) { printf(“\nOut of memory!\n”); exit(1); } ... free(rect); Joint struct definition and typedef Check for returned NULL Do NOT change rect during this time!!! Summer 2012 -- Lecture #4

  20. Question: Want output: a[] = {0,1,2} with no errors. Which lines do we need to change? 1 #define N 3 2int *makeArray(int n) { 3int *ar; 4ar = (int *) malloc(n); 5 return ar; 6 } 7 void main() { 8inti,*a = makeArray(N); 9 for(i=0; i<N; i++) 10 *a++ = i; 11printf(“a[] = {%i,%i,%i}”,a[0],a[1],a[2]); 12 free(a); 13 } 4, 12 ☐ 5, 12 ☐ 4, 10 ☐ 5, 10 ☐

  21. Agenda • C Memory Layout • Stack, Static Data, and Code • Administrivia • Dynamic Memory Allocation • Heap • Memory Management • C Wrap-up: Linked List Example • Bonus: Common Memory Problems Summer 2012 -- Lecture #4

  22. Memory Management • Many calls to malloc() and free() with many different size blocks – where are they placed? • Want system to be fast with minimal memory overhead • Versus automatic garbage collection of Java • Want to avoid fragmentation, the tendency of free space on the heap to get separated into small chunks Summer 2012 -- Lecture #4

  23. B2 (1 byte) Fragmentation Example . . . B4 • Block 1: malloc(100) • Block 2: malloc(1) • Block 1: free(B1) • Block 3: malloc(50) • What if malloc(101)? • Block 4: malloc(60) B3? Fragment B1 (100 bytes) B3 B3? Fragment Heap Summer 2012 -- Lecture #4

  24. Basic Allocation Strategy: K&R • Section 8.7 offers an implementation of memory management (linked list of free blocks) • If you can decipher the code, you’re well-versed in C! • This is just one of many possible memory management algorithms • Just to give you a taste • No single best approach for every application Summer 2012 -- Lecture #4

  25. K&R Implementation • Each block holds its own size and pointer to next block • free() adds block to the list, combines with adjacent free blocks • malloc() searches free list for block large enough to meet request • If multiple blocks fit request, which one do we use? Summer 2012 -- Lecture #4

  26. Choosing a Block in malloc() • Best-fit: Choose smallest block that fits request • Tries to limit wasted fragmentation space, but takes more time and leaves lots of small blocks • First-fit: Choose first block that is large enough (always starts from beginning) • Fast but tends to concentrate small blocks at beginning • Next-fit: Like first-fit, but resume search from where we last left off • Fast and does not concentrate small blocks at front Summer 2012 -- Lecture #4

  27. Question: Which allocation system and set ofrequests will create a continuous block in the Heap?B3 was the last fulfilled request. Best-fit:malloc(50), malloc(50) ☐ 50 10 B2 30 First-fit: malloc(50), malloc(30) ☐ 10 B3 Next-fit:malloc(30), malloc(50) ☐ 50 B1 Next-fit:malloc(50), malloc(30) ☐ 50

  28. Get To Know Your Instructor Summer 2012 -- Lecture #4

  29. Agenda • C Memory Layout • Stack, Static Data, and Code • Administrivia • Dynamic Memory Allocation • Heap • Memory Management • C Wrap-up: Linked List Example • Bonus: Common Memory Problems Summer 2012 -- Lecture #4

  30. Linked List Example • We want to generate a linked list of strings • This example uses structs, pointers, malloc(), and free() • Create a structure for nodes of the list: struct Node { char *value; struct Node *next; }; Notice the recursive definition! Summer 2012 -- Lecture #4

  31. Simplify Code with typedef • It gets annoying to type out structListNode over and over again • Define new variable type for struct: Method 1:Method 2: • Can further rename pointers: typedefListNode * List; typedefchar * String; struct Node { char *value; struct Node *next;}; typedefstruct Node ListNode; typedefstruct Node { char *value; struct Node *next;} NodeStruct; List myLinkedList; String value; Summer 2012 -- Lecture #4

  32. Adding a Node to the List In what part of memory are these stored? • Want functionality as shown: String s1 = "start", s2 = "middle"; String s3 = "end"; List theList = NULL; theList = addNode(s3, theList); theList= addNode(s2, theList); theList= addNode(s1, theList); Must be able to handle a NULL input If you’re more familiar with Lisp/Scheme, you could name this function cons instead. Summer 2012 -- Lecture #4

  33. Adding a Node to the List • Let’s examine the 3rd call ("start"): List addNode(String s, List list) { List node = (List) malloc(sizeof(NodeStruct)); node->value = (String) malloc (strlen(s) + 1); strcpy(node->value, s); node->next = list; return node; } Don’t forget this for the null terminator! "middle" "end" "start" "start" ??? node: list: s: NULL ? ? Summer 2012 -- Lecture #4

  34. Removing a Node from the List • Delete/free the first node ("start"): List deleteNode(List list) { List temp = list->next; free(list->value); free(list); return temp; } What happens if you do these in the wrong order? temp: "start" ??? "middle" "end" list: NULL ??? Summer 2012 -- Lecture #4

  35. Additional Functionality • How might you implement the following: • Append node to end of a list • Delete/free an entire list • Join two lists together • Reorder a list alphabetically (sort) Summer 2012 -- Lecture #4

  36. Summary • C Memory Layout • Static Data:globals and string literals • Code: copy of machine code • Stack: local variables (grows & shrinks in LIFO manner) • Heap: dynamic storage using mallocand freeThe source of most memory bugs! • Memory Management • Want fast with little fragmentation Summer 2012 -- Lecture #4

  37. BONUS SLIDES You are responsible for the material contained on the following slides, though we may not have enough time to get to them in lecture. They have been prepared in a way that should be easily readable and the material will be touched upon in the following lecture. Summer 2012 -- Lecture #4

  38. Agenda • C Memory Layout • Stack, Static Data, and Code • Administrivia • Dynamic Memory Allocation • Heap • Memory Management • C Wrap-up: Linked List Example • Bonus: Common Memory Problems Summer 2012 -- Lecture #4

  39. Know Your Memory Errors (Definitions taken from http://www.hyperdictionary.com) Less common in 61C More common in 61C • Segmentation Fault “An error in which a running Unix program attempts to access memory not allocated to it and terminates with a segmentation violation error and usually a core dump.” • Bus Error “A fatal failure in the execution of a machine language instruction resulting from the processor detecting an anomalous condition on its bus. Such conditions include invalid address alignment (accessing a multi-byte number at an odd address), accessing a physical address that does not correspond to any device, or some other device-specific hardware error.” • Segmentation Fault “An error in which a running Unix program attempts to access memory not allocated to it and terminates with a segmentation violation error and usually a core dump.” • Bus Error “A fatal failure in the execution of a machine language instruction resulting from the processor detecting an anomalous condition on its bus. Such conditions include invalid address alignment (accessing a multi-byte number at an odd address), accessing a physical address that does not correspond to any device, or some other device-specific hardware error.” Summer 2012 -- Lecture #4

  40. Common Memory Problems • Using uninitialized values • Using memory that you don’t own • Using NULL or garbage data as a pointer • De-allocated stack or heap variable • Out of bounds reference to stack or heap array • Using memory you haven’t allocated • Freeing invalid memory • Memory leaks Summer 2012 -- Lecture #4

  41. Using Uninitialized Values • What is wrong with this code? void foo(int *p) { int j; *p = j; } void bar() { inti=10; foo(&i); printf("i = %d\n", i); } j is uninitialized (garbage), copied into *p Using i which now contains garbage Summer 2012 -- Lecture #4

  42. Using Memory You Don’t Own (1) • What is wrong with this code? typedefstruct node { struct node* next; intval; } Node; intfindLastNodeValue(Node* head) { while (head->next != NULL) head = head->next; return head->val; } What if headis NULL? No warnings! Just Seg Faultthat needs finding! Summer 2012 -- Lecture #4

  43. Using Memory You Don’t Own (2) • What’s wrong with this code? char *append(const char* s1, const char *s2) { const int MAXSIZE = 128; char result[MAXSIZE]; inti=0, j=0; for (j=0; i<MAXSIZE-1 && j<strlen(s1); i++,j++) result[i] = s1[j]; for (j=0; i<MAXSIZE-1 && j<strlen(s2); i++,j++) result[i] = s2[j]; result[++i] = '\0'; return result; } Local array appears on Stack Pointer to Stack (array) no longer valid once function returns Summer 2012 -- Lecture #4

  44. Using Memory You Don’t Own (3) • What is wrong with this code? typedefstruct { char *name; int age; } Profile; Profile *person =(Profile *)malloc(sizeof(Profile)); char *name = getName(); person.name = malloc(sizeof(char)*strlen(name)); strcpy(person.name,name); ... // Do stuff (that isn’t buggy) free(person); free(person.name); Did not allocate space for the null terminator! Want (strlen(name)+1) here. Accessing memory after you’ve freed it. These statements should be switched. Summer 2012 -- Lecture #4

  45. Using Memory You Haven’t Allocated • What is wrong with this code? void StringManipulate() { const char *name = “Safety Critical"; char *str = malloc(10); strncpy(str, name, 10); str[10] = '\0'; printf("%s\n", str); } Write beyond array bounds Read beyond array bounds Summer 2012 -- Lecture #4

  46. Using Memory You Haven’t Allocated • What is wrong with this code? char buffer[1024]; /* global */ int main(intargc,char *argv[]) { strcpy(buffer,argv[1]); ... } What if more than a kibi characters? This is called BUFFER OVERRUN or BUFFER OVERFLOW and is a major security flaw!!! Summer 2012 -- Lecture #4

  47. Freeing Invalid Memory • What is wrong with this code? void FreeMemX() { intfnh = 0; free(&fnh); } void FreeMemY() { int *fum = malloc(4*sizeof(int)); free(fum+1); free(fum); free(fum); } 1) Free of a Stack variable 2) Free of middle of block 3) Free of already freed block Summer 2012 -- Lecture #4

  48. Memory Leaks • What is wrong with this code? int *pi; void foo() { pi = (int*)malloc(8*sizeof(int)); ... free(pi); } void main() { pi = (int*)malloc(4*sizeof(int)); foo(); } Overrode old pointer! No way to free those 4*sizeof(int) bytes now foo() leaks memory Summer 2012 -- Lecture #4

  49. Memory Leaks • Rule of Thumb: More mallocs than frees probably indicates a memory leak • Potential memory leak: Changing pointer – do you still have copy to use with free later? plk = (int *)malloc(2*sizeof(int)); ... plk++; Typically happens through incrementation or reassignment Summer 2012 -- Lecture #4

  50. Debugging Tools • Runtime analysis tools for finding memory errors • Dynamic analysis tool: Collects information on memory management while program runs • No tool is guaranteed to find ALL memory bugs; this is a very challenging programming language research problem • You will be introduced to Valgrind in Lab 3 http://valgrind.org Summer 2012 -- Lecture #4

More Related